What will fill those fabs?
I'll cover fab capacity first, and then talk a bit about products...
Since we are used to looking at CPU's we have a bit of a skewed perspective on what a fab can do. A typical INTC CPU factory cranks out about $7-8B/year in revenue, so Intel needs about 7 or 8 factories making "other" products to equal one factory making CPU's... again if you use the 2X multiplier for 300mm you still get around 4 "other" factories per cpu factory.
Today, for 90nm Intel has D1C, F11X and F24? D1C converts to 65nm leaving only 2 90nm factories for "other" to soak up. Using our basic algorithms these two factories would produce about $4B/year in revenue... leaving quite a hole for the rest of the network to soak up. Intel has announced the new low power 65nm process. I suspect of the four 65nm factories about 1.5 of them will be used for lower power "other" stuff and about 2.5 of them for CPU's. Intel has yet to announce where 32nm development will be done. I suspect current development is in RP1, but it will need a home. 45nm capacity will likely be upgrades to older factories, as such it is likely that next year INTC will take a few older factories off line to begin the 45nm conversion.
In terms of products, there will be mobile chipsets on the low power 65nm process. Desktop chipsets will move to 90nm. I suspect Xscale will move to the low power 65nm process. If you were to buy a PDA today (a small market though) it would contain an Intel Xscale CPU. Intel has continued to get design wins in the cell phone space, and this market can certainly consume lots of silicon. With the new iAMT I suspect the Intel share of ethernet controllers is going to rise a bit. The original centrino radio was manufactured by Philips, but the current version comes from an INTC factory.
--Alan
News 
Market Data 
Discover 
AI
